Unify S5-nano model cards with an introductory format paragraph

README.md

@@ -8,6 +8,8 @@ library_name: transformers
 
 # S5-nano
 
+This repository contains the **BF16 version of S5-nano** (8.7 GB), with the original, unquantized weights for vLLM and SGLang. The [FP8 version](https://huggingface.co/maritaca-ai/S5-nano-FP8) reduces the weight size to 5.7 GB and is recommended for GPUs with native FP8 support. The [GGUF versions](https://huggingface.co/maritaca-ai/S5-nano-GGUF) are for llama.cpp and compatible applications, with BF16 (8.7 GB), Q8_0 (4.6 GB), and Q4_K_M (2.8 GB) options.
+
 S5-nano is a 4B-parameter model from Maritaca AI, derived from
 [Qwen3.5-4B](https://huggingface.co/Qwen/Qwen3.5-4B). It has two modes:
 
@@ -28,6 +30,11 @@ The context window supports up to 262,144 tokens, and the weights include an MTP
 FP8 is recommended for GPUs with native FP8 support (Ada, Hopper, and Blackwell).
 The GGUF repository provides BF16, Q8_0, and Q4_K_M variants.
 
+The FP8 version uses E4M3 weights for linear layers, with one scale per 128×128
+block, and dynamically quantizes activations during inference. Embeddings, the
+vision encoder, and some small projections remain in BF16. The MTP layer is
+included.
+
 <!-- s5-nano-evaluation-results:start -->
 ## Evaluation results
 
@@ -962,6 +969,9 @@ Optional flags:
 For the FP8 version, replace the model name with `maritaca-ai/S5-nano-FP8`
 and keep the same flags.
 
+When serving FP8, also use `maritaca-ai/S5-nano-FP8` in the request's
+`model` field (or `MODEL` in the Python examples).
+
 ### Requests with curl
 
 Think mode: